Asking how improbable the data would be if there were no effect — the frame in which discoveries are declared.Null hypothesis testing is a statistical framework used to decide whether an observed effect in data is real or could plausibly be due to random chance. The 'null hypothesis' typically states that there is no effect, no difference, or no relationship — for example, that a new drug works no better than a placebo. The test asks: if the null hypothesis were true, how unlikely would it be to see data at least as extreme as what was actually observed? If that unlikelihood falls below a pre-set threshold (commonly 5%), researchers declare the result 'statistically significant' and reject the null hypothesis in favor of an alternative.
The framework was formalized in the early 20th century by Ronald Fisher, who introduced the concept of significance testing in the 1920s, and later extended by Jerzy Neyman and Egon Pearson with the alternative hypothesis and error types.
The procedure begins by assuming the null hypothesis is true. From this assumption, one derives a probability distribution for a test statistic — a number summarizing the data, like a difference in means. The observed data yields an actual value of that statistic. The test then computes the probability of obtaining a value as extreme or more extreme than the observed one, under the assumed distribution. This probability is the p-value. If the p-value is small enough, the result is deemed inconsistent with the null hypothesis, and the null is rejected.
It is the default tool in scientific research, from clinical trials to psychology experiments, for deciding whether an intervention or effect is worth attention.
The 5% threshold for significance was popularized by Fisher, who chose it as a convenient round number in his 1925 book 'Statistical Methods for Research Workers'.












